Real time datawarehouse replacing ooperational reporting

Hi,

As I was reading through articles I came across one of the advantages in real time datawarehouse that it can replace the operational reporting so there is no need for reports to go across 2 reporting DB there can be one centralised Real time DW so that operational and analytical reports can be generated in the same area.

Any thoughts arounds this??? Just wanted to have the your idea onthis can Real timeDW reporting replace operational?? how does it sound cost wise?

Sounds like a sales pitch. OLTP is real time. That's where I go when I want real time.

Well but dont you think it can save cost provided the functionality is provided so that business does not have to go to 2 places to do reporting and imagine data redundancy as operational may require replica datastore and the total cost that an enterprise can save. But the questions is real time dw achievable and how much is it going to cost the enterprise, needs for realtime DW all these things should be considered right?

Well but dont you think it can save cost provided the functionality is provided so that business does not have to go to 2 places to do reporting and imagine data redundancy as operational may require replica datastore and the total cost that an enterprise can save

Thats the problem... so called 'real time DW' doesn't change that.

First, the operational system needs it data, otherwise it wouldn't be very operational. Second, the data model to support operational reporting is different than a data model to support analytic reporting. The operational data model is typically a near-3NF model... it is often refered to as an ODS, or operational data store, and it typically used to support small specific queries against document centric/transactional data. An analytic data warehouse is best designed using dimensional models as analytic work usually involves aggregate queries against large amounts of data.

Properly implented, supporting operational requirements in a data warehouse requires greater data redundancy and significantly more complex processes.

Properly implented, supporting operational requirements in a data warehouse requires greater data redundancy and significantly more complex processes.

Well Operational reporting usually is very simple and unless the adhoc reporting or score card/dashboard has completely different look and feel for operational as opposed to analytical, but though the model is different 3NF to dimensional how does that change the perspective of the user getting the appropriate real time results based on operational data when data is fed into datawarehouse real time. One thing that could be challenge from Operational to Datawarehouse is High availability of the operational reporting to analytical reporting. DATA REDUNDANCY is what is confusing me how can data be redundant if the same data can be used for analytical and operational purpose.

This is an interesting article

http://www.informationweek.com/news/global-cio/showArticle.jhtml?articleID=167100483

I am just trying to get an expert view of all the available options and explore them..

That means Overstock needed a high-volume data-management tool to migrate data into Oracle 10g and then load selected data into the new NCR Corp. Teradata data warehouse,

... from the article. Notice the data is landing into two places. The real-time operational stuff first goes to 10g to support operational reporting followed by 'selected' data going to Teradata to support analytics. They are using GoldenGate to get the data into the 10g environment, which is basically, doing data replication.

The reduncancy comes from the need to create different data models to EFFICIENTLY handle different reporting purposes (operational versus analytic). You could do analytic reporting against a 3NF data model (even Inmon advises against it), if your users are willing to tollerate multi-hour response times on reports.